高阶时滞对象的预测PI(D)控制  

Predictive PI(D) control for high order processes with time-delay

作　　者：任正云[1] 张红[2] 邵惠鹤[3] 

机构地区：[1]东华大学自动化系,上海200051 [2]上海工程技术大学化工学院,上海200065 [3]上海交通大学自动化系,上海200030

出　　处：《控制理论与应用》2005年第4期645-648,652,共5页Control Theory & Applications

基　　金：国家十五863项目资助课题(2001-AA413130).

摘　　要：利用频率域模型降阶理论,提出了高阶时滞对象的预测PI(D)控制器两种设计方法.一种方法是直接将高阶滞后对象在频率域内降阶为低阶滞后对象,针对低阶滞后对象设计预测PI(D)控制器;另一种方法是按照规定的性能指标设计控制器,并将该控制器在频率域内降阶为具有预测PI(D)控制器的结构形式.这两种方法设计的控制器均具有结构简单、可调参数少、参数调节方便的特点.仿真表明:在模型失配的情况下,此两类预测PI(D)控制器仍然具有良好的控制性能和鲁棒稳定性能.Based on the model reduction theory in frequency-domain, two predictive PI（D） controller design methods for high order processes with time-delay are proposed.The first method is that the predictive PI（D） controller is designed by a low order model reduced directly from the process model. The second one is that a controller is first designed by the prescribed performance index, and then the controller is simplified to a predictive PI（D） controller. The advantages of these controllers are simple structure, small number of tuning parameters and easy tuning procedures. Illustrative examples are offered to show the good performance and robust stability of proposed schemes under model-process mismatch.

关 键 词：预测PI(D) 高阶时滞对象 模型降阶 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]